Aerodynamic Trim Conditions 2

[RoopB]

Hello all,

I have a program which requires input such as Velocity, Angle of attack, and Fin deflection.

The program goes through and calculates the aerodynamic coefficients for my vehicle, then outputs my Forces and Moments on the body (Fx, Fy, Fz, Mx, My, Mz).

I need to make a "wrapper" of sorts that can find the trim conditions of my Vehicle. In other words, I would like to set my Moments all equal to zero, and have the trim wrapper iterate until it finds what fin deflection will give M = 0.

Is it possible to do some sort of numerical analysis in Matlab to do this? I am not sure where to start, is there a paticular method I should be looking at to do this?

Thanks in advance for your thoughts.

 

[FeX32]

Are you familiar with numerical optimization?
This is ideal for any sort of parametric determination/ determining conditions. ect.


Fe

 

[GregLocock]

Do you have the Optimization Toolbox?

If not you need to set up a function that gives (Mx^2+My^2+Mz^2) as a result of changes in Fin deflection, then use whatever root finding recipe takes your fancy.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[RoopB]

Thanks,

I'm not faimiliar with numerical optimization (yet). I do have the optimization toolbox in matlab, and I'm in the process of trying to determine the best way to attain my goal. I have looked at the fsolve and fgoalattain functions, but I'm not sure that they are able to do what I need. I rewrote my program to be a function in the form [Fx, Fy, Fz, Mx, My, Mz] = AeroFunc(Velocity, AOA, Phi, findeflection)

I have multiple inputs and multiple outputs, do you think there is a way to set my Mx, My, Mz to zero, with a specified Velocity, Angle of attack, and Phi, while letting the function solve for findeflection? I'm afraid I'm asking too much. Is there a function that can do that?

Thanks again.

 

[FeX32]

If I understand you correctly, you can determine a particular solution with a multi-objective cost function and some constraints.

"Mx, My, Mz to zero" can be an objective. ie. minimize the absolute value of the M's.

"a specified Velocity, Angle of attack, and Phi"
This means that these are now constant, not playing a role in the objective optimization process.

"while letting the function solve for findeflection"
From your function it looks like this is an input to your function.
Likely, this can act as your (or one of your) design variable.

fsolve is a pretty good one to start with. I have used it before with success. Watch out for the initial point, it can cause some discrepancies as sometimes when too far it causes fsolve to converge to a local min.

Personally, I ended up writing my own optimization codes, sometimes starting by modifying other standard ones.
I don't suggest you go this route unless you are proficient in coding as well as very familiar with optimization procedures.
The opti. toolbox is more then useful for 99% of things.


Fe